I have three graduation caps thrown in the air that interact with each other. I fed them to an RBD sop solver with different attributes like velocity and w for each of them to differentiate their behavior. Now, I need each of them to have a tassel ribbon attached to it that flaps around as they fly. I did tests with vellum cloth and pin to animation. This works if I do this to one single cap that has been animated by hand. However, now I have three caps that come out of an RBD sop solver. How do I approach this, considering that I should preserve the current RBD sim as the caps now fly exactly as I need them to? The tassel is part of the whole cap object, and I separated it with groups and blasts.

Hi, I'm kind of stuck with something I'm trying to achieve. If anybody could point me in the right direction, that would be fantastic. I'm trying to combine two techniques in the same shot.

I have an image that drives a flip simulation. Image has two colors and each color will become a different fluid that interacts/mixes with the other one. Initially, this image was still and it was easy for me to achieve that with attributesfrommap and mapping Cd to viscosity and density.

Now I need the underlying image to be an animation (only the character moves, the rest is still). I got the character animation driving the particles representing its shape using the volume optical flow technique. However, I'm struggling to find a way to merge these two systems together and basically add the character movement to the fluids mixing together generated from image Cd.

Attached is a test project folder, which includes the images to drive the sim.

Thanks.

Hi, how do you use two separate particle fluid surfaces from one single DOP sim? I have a sim with two fluids interacting with one another. I used attributes to distinguish them in terms of density, viscosity, and color. However, when I render, I need the cream fluid to slowly emerge from coffee and see through the coffee as it does so. Keeping it in one geo I can't do it as there is no cream geo under the coffee geo and by simply using the Cd attribute, I can't see the cream until it is on the surface. Also, the coffee geo is flattened to attach it to an ocean system. Here is my nodegraph. Any help would be great. Thanks!

Hi,

I'm trying to achieve an effect where a creamer gets poured into a coffee cup and the two fluids mix nicely together and smoothly like in the linked video [vimeo.com].

I'm attaching my project file. So far I either get it too splashy or too dense where there is no nice mixing. If anybody could point me in the right direction that would be super appreciated.

Thanks!

Hi, Houdini won’t launch unless I manually launch hserver before. The problem started suddenly and I didn’t install Houdini updates or changed anything I can think of in my system before the problem started to occur.

I re-installed Houdini 19 and hserver but the problem persists. During the installation I’m prompted to run the installer as root and enter the password so I don’t think that’s the problem.

Any idea how to fix this? Is hserver supposed to start when you launch Houdini or at boot?

I’m on Linux Ubuntu 20.04 - Houdini 19 py2.

Thanks.

UPDATE: it now works - I was caching the VDB in the wrong way. However, I hope they will fix this problem soon and allow us to use SOPimport for volumes to be rendered with Arnold in Solaris. It is very limiting to have to write VDB's every time if one wants to build a procedural system.

I'm attaching the file in case anyone needs help with this.

Hi,
I'm unable to render volumes in Solaris with Arnold. I read that sopImport doesn't work with volumes for Arnold so I exported as VDB and imported in lops with both volume and file nodes. The scene renders fine if I use Karma but if switch to Arnold the VDB won't render at all. I'm attaching a sample scene. Can anybody take a look and let me know what I'm doing wrong. Thanks!

UPDATE: I found the solution and I thought to share it in case anybody else had this problem. The issue was in the Points from Volume node that I used to generate the points to feed to the POP solver. Changing the construction method from Sparse Volume to Dense Grid removed those leaking extra points that created the flicker.

Hi,

I am working on a scene where I have a particle system that forms gradually. There is a barrel in the scene that progressively dissolves, and a foam-like substance takes its place before melting down. The barrel animation is driven by a pyro source spread, that in conjunction with an Axiom solver, also drives the velocity of the POP simulation. The simulation works fine. However, in rendering (with Arnold), I have many tiny particles under the intended particles that reflect the light source while flickering, and these are not present on every frame. This flicker only appears in the render. I cannot see anything between the "real" particles and the box object below in the viewport. I am attaching the first few frames + an annotated frame. Any idea what is causing my issue? Thanks

Herve
Hi,
As som said, in such cases it's essential to keep orthogonal vectors between ‘up’ and ‘N’.
Since you already set up your polyframe correctly, I'd say that you just need to adjust your “final” backbone path N attribute in Y only (presupposing the usual default of up being 0,1,0 and your pillar Y effectively being their verticals). Unless I'm missing something… but try it with an attrib create, or wrangle, on points.

I added a pic to hopefully clarify my non-native verbosity…

this will keep your N.y flat, so the pillars will not bank anymore. Now if your road itself would bank, you might get some clipping again, so… just FYI.

This worked. Thank you!

som
Are you generating an ‘up’ as well? That helps to align the pillars. Make sure the normal up and tangents are orthogonal to eachother.

Thanks for the reply. Could you please make an example? Sorry - quite a beginner here. I understand the theory of what you are saying but I cannot figure out how to do it in practice.

So far I generated the normals for that copy to points with polyframe, setting the Tangent name to N. Then I reversed them with @N = -@N; This is how I got to this point as depicted in the attached screenshot.

Hi, I have a procedural road that follows any given curve. The road is generated by a simple sweep of two lines.

Now, the road has base pillars under it. These are copied to points belonging to the same curve that generates the road direction/length.

When the road curves the pillars's ends protrude from the road. I believe this is because the pillars are aligned with the points normals of the curve that is at the center of the road and these don't consider the actual generated road's normals.

How do I fix this?

Thanks

xilofoton
Beyond Dneg, is there any place where it's still the part of the pipeline?
No FMX and subtle Siggraph 2019 presence by Iso…
Things may changed, I think we have better options, like RfH, Pixar just released the second iteration of the rewritten bridge, 22.6.
It seems a Pseudoclarisse engine for Houdini with the live feedback (and more versatile ie live displacement), with Pixar robustness, but haven't try it yet…
https://rmanwiki.pixar.com/display/RFH22/RenderMan+for+Houdini+22.6 [rmanwiki.pixar.com]

Clarisse and Houdini are very different tools, you can’t compare them. I use and love both but I use them for different things. I create assets procedurally in Houdini and use Clarisse for lookdev, layout and rendering.

Although instancing technologies are becoming more popular with the various renderers, I still think that Clarisse is unique in terms of dealing with massive scenes such as mega environments, which is the stuff that I do. At the moment I’m working on a scene in Clarisse that has trillions of polygons, hundreds of UDIM textures, hundreds of regular 4K textures, and I can still work in the viewport in almost real-time and rendering is pretty fast as well.

I have tried Renderman 22.6 for Houdini and was great and have also been using RS. They are great tools but I don’t think you could handle one of those massive environments you can do in Clarisse inside Houdini, at least for the time being (H18 will probably make it possible).

That being said, Houdini engine for Clarisse would make sense to me as it would be incredibly powerful to create variations of the same asset directly inside Clarisse at scattering time.

I found a generic car rig by SideFX on Orbolt at this https://www.orbolt.com/asset/SideFX::vehiclerig

This seems very useful but it doesn't seem to be working (maybe because it's an old asset?). Does anybody have any information about this?

I also tried the newer Compact Car HDA, which appeared to be working (but I only tested it for a few minutes). However, this one is not intended to be used with your own car model I believe?

Thanks

Thanks for your inputs. I ended up doing a variation of the sort + transform suggestion. It now works.

Hi,

Sorry for the silly question but I'm stuck. It's hard to describe what I want to do with words so I'm attaching an annotated screenshot.

Basically, I need to translate procedurally the wire in the picture so that the top point (A) is always at 0,0. The closest I could get is using the expressions you can also see in the picture. Y alignment works but X alignment is offset because if use bbox D_XMIN the “left-most” point (B) is not on top but around the middle of the wire. Hope this makes sense. Thanks!

+1 Houdini Engine for Clarisse

aRtye
You can. The issue is OsX adds a .local to the hostname that Windows refuses to add. So the combos that work are:

OsX / Linux
Linux / Windows

I actually got this to work on the combo MacOS Sierra/Windows 10 Pro. I had to do a lot of messing around to make this work so I'm not even sure what was the key step that fixed the problem above described by aRtye. The theory behind it is that .local is not actually part of the hostname but it's rather the domain where the hostname is found. For what I understand, MacOS uses .local by default, Win doesn't.

In the system settings in windows I selected to use a domain (I believe in place of Workgroup)and typed in .local as domain. I'm not sure anymore where exactly I typed this in but eventually windows would start listing a “Full computer name” which adds .local to it (of course the rest of the name has to match the MacOS computer name).

I also changed the server name on the licensing app to the new Full computer name. Basically you can click the “local” button; this will put in the name without .local and then you manually add .local at the end of the name.

Also, still in the licensing app, I had to reinstall the server key (there is an option for that).

Once all this is done you should be able to add your Indie license to Windows.

This worked in the scenario where I had my Indie license already up and running on MacOS and later on I installed this on my secondary Bootcamp partition with Win10. I don't know if this works the other way around.

Hope this helps.

Hello,

I understand that Houdini Indie will send the scene information directly to the third party renderer without intermediate steps. Will this prevent to render using such third-party renderers on a pay-per-minute remote cloud farm service?

I'm considering to switch to Houdini after I read about the third party rendering support in Indie but I would need to be able to send my render jobs to cloud services. Is this possible?

Thanks

Hello,

I read that Houdini Indie will send the scene information directly to the third party renderer without intermediate steps. Will this prevent to render using such third-party renderers on a pay-per-minute remote cloud farm service?

I'm considering to switch to Houdini after I read about the third party rendering support in Indie but I would need to be able to send my render jobs to cloud services.

Thanks